Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{198470,
        author = {K.K.B. Vali Bhasha SK and Hafsa Mehek Pathan and K Siva Krishna and R.Guru Charan and Lakkireddy Venkata Satish},
        title = {Anomaly Detection in Network Traffic Using Advanced Machine Learning Techniques},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {10230-10238},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=198470},
        abstract = {The rapid growth of Internet of Things networks has brought about numerous security issues due to their distributed nature, computational limitations, and susceptibility to a wide range of cyberattacks. Traditional intrusion detection techniques, which were primarily based on static and wired infrastructures, are not adequate to deal with the dynamic nature of contemporary cyberattacks. Therefore, in order to overcome the drawbacks of traditional intrusion detection techniques, an intelligent intrusion detection system based on deep reinforcement learning will be proposed in this paper. In this context, the intrusion detection system will be modeled based on the Markov Decision Process, wherein the system states will be defined based on the network flow feature, and the actions will be defined based on the intrusion detection decisions, while the reward mechanism will be defined based on the trade-off between accuracy, false alarms, and response efficiency. The proposed framework is evaluated by employing the UNSW_NB15 dataset in a simulated network environment. Two sophisticated reinforcement learning models, namely Deep Deterministic Policy Gradient and Proximal Policy Optimization, are implemented for the detection and classification of multiple attack types, including distributed denial of service, denial of service, backdoor, injection, man in the middle, password attacks, ransomware, and cross site scripting. The experimental results reveal that the proposed Deep Deterministic Policy Gradient model has an accuracy of 98 percent, whereas the proposed Proximal Policy Optimization model has an accuracy of 89.90 percent, thereby proving the effectiveness and adaptability of the proposed system.},
        keywords = {Internet of Things security, Intrusion detection system, Deep reinforcement learning, Deep Deterministic Policy Gradient, Proximal Policy Optimization and Network attack detection.},
        month = {April},
        }